The present invention discloses a delay request processing method and apparatus that are based on a token bucket and relates to the field of communications technologies, where the method and the apparatus are invented to shorten a scanning period when a delay request is processed. The method includes: receiving a delay request message of a queue corresponding to a token bucket, where the delay request message includes delay request time; determining, according to the delay request time, whether the delay request message needs to be processed preferentially; processing the delay request message in a preferentially processing manner when it is determined that the delay request message needs to be processed preferentially. The present invention may be applied to a delay request processing technology.